Nitric Oxide Scavenging by Red Blood Cells as a Function of Hematocrit and Oxygenation*

The reaction rate between nitric oxide and intraerythrocytic hemoglobin plays a major role in nitric oxide bioavailability and modulates homeostatic vascular function. It has previously been demonstrated that the encapsulation of hemoglobin in red blood cells restricts its ability to scavenge nitric oxide. This effect has been attributed to either factors intrinsic to the red blood cell such as a physical membrane barrier or factors external to the red blood cell such as the formation of an unstirred layer around the cell. We have performed measurements of the uptake rate of nitric oxide by red blood cells under oxygenated and deoxygenated conditions at different hematocrit percentages. Our studies include stopped-flow measurements where both the unstirred layer and physical barrier potentially participate, as well as competition experiments where the potential contribution of the unstirred layer is limited. We find that deoxygenated erythrocytes scavenge nitric oxide faster than oxygenated cells and that the rate of nitric oxide scavenging for oxygenated red blood cells increases as the hematocrit is raised from 15% to 50%. Our results 1) confirm the critical biological phenomenon that hemoglobin compartmentalization within the erythrocyte reduces reaction rates with nitric oxide, 2) show that extra-erythocytic diffusional barriers mediate most of this effect, and 3) provide novel evidence that an oxygen-dependent intrinsic property of the red blood cell contributes to this barrier activity, albeit to a lesser extent. These observations may have important physiological implications within the microvasculature and for pathophysiological disruption of nitric oxide homeostasis in diseases.

The reaction rate between nitric oxide and intraerythrocytic hemoglobin plays a major role in nitric oxide bioavailability and modulates homeostatic vascular function. It has previously been demonstrated that the encapsulation of hemoglobin in red blood cells restricts its ability to scavenge nitric oxide. This effect has been attributed to either factors intrinsic to the red blood cell such as a physical membrane barrier or factors external to the red blood cell such as the formation of an unstirred layer around the cell. We have performed measurements of the uptake rate of nitric oxide by red blood cells under oxygenated and deoxygenated conditions at different hematocrit percentages. Our studies include stopped-flow measurements where both the unstirred layer and physical barrier potentially participate, as well as competition experiments where the potential contribution of the unstirred layer is limited. We find that deoxygenated erythrocytes scavenge nitric oxide faster than oxygenated cells and that the rate of nitric oxide scavenging for oxygenated red blood cells increases as the hematocrit is raised from 15% to 50%. Our results 1) confirm the critical biological phenomenon that hemoglobin compartmentalization within the erythrocyte reduces reaction rates with nitric oxide, 2) show that extraerythocytic diffusional barriers mediate most of this effect, and 3) provide novel evidence that an oxygen-dependent intrinsic property of the red blood cell contributes to this barrier activity, albeit to a lesser extent. These observations may have important physiological implications within the microvasculature and for pathophysiological disruption of nitric oxide homeostasis in diseases.
Nitric oxide (NO) 3 is an endothelium-derived relaxation factor that is synthesized in endothelial cells (1)(2)(3)(4). To elicit its vasodilatory activity, NO must diffuse to the smooth muscle cells and activate soluble guanylate cyclase. In 1994, Lancaster suggested that the proximity of the endothelium to the millimolar concentrations of hemoglobin (Hb), an avid NO scavenger, would severely compromise the efficiency of the NO/soluble guanylate cyclase pathway (5). However, later studies have indicated that the physical compartmentalization of hemoglobin within the red blood cell (RBC) effectively reduces the apparent rate at which NO is consumed by Hb (6 -15). One contributory element to this effect is a RBC-free zone at the blood/endothelium interface that is present during laminar flow (7,9,10). In addition, the rate of NO consumption has been reported to occur up to 1000 times more slowly by red blood cells than by an equivalent concentration of cell-free hemoglobin. Two potential mechanisms for this effect involve either the presence of an unstirred layer surrounding the red blood cell that is formed as a result of NO diffusion (6,13) or a physical barrier to NO diffusion that is integral to the protein-rich RBC submembrane (11). The faster effective reaction of NO with cell-free Hb compared with RBC-encapsulated hemoglobin may partially explain the hypertensive effects of some hemoglobin-based blood substitutes (16 -22) and contribute to the vascular dysfunction associated with hemolytic anemias such as sickle cell disease (23,24).
The contribution of each of the RBC-associated factors (unstirred layer versus an intrinsic membrane barrier) to reduced NO scavenging in blood is the subject of some debate and is a major issue addressed in this study (11,13,15). Rapid mixing of NO with RBCs directly measures the rate of NO uptake but both the unstirred layer and intrinsic membrane barrier could potentially be responsible for the slow NO uptake observed in these measurements. In 2000, Liao and coworkers (11) developed a technique to measure the rate of NO uptake by RBCs in which the potential contribution of the unstirred layer is minimized. This technique uses competition experiments that involve slow release of NO in the presence of excess cell-free and RBC-encapsulated Hb. The rate of NO uptake by the RBCs is calculated using the known rate that NO reacts with cell-free Hb and measuring the fraction of reacted Hb in the RBCs compared with that in solution (11). Using this competition method, Liao and coworkers found that the rate that NO is taken up by oxygenated RBCs at 15% hematocrit (Hct) is 500 -1000 times slower than the rate that cell-free Hb binds NO, implicating a major role for the intrinsic membrane boundary (11).
Recent calculations have found that the apparent rate of NO scavenging by RBCs in competition experiments can be explained without incorporation of an intrinsic membrane boundary due to its dependence on the bimolecular rate of the reaction of free Hb with NO and on Hct (25). We have examined these factors in this report. To better understand the contribution of extracellular diffusional barriers, the potential intrinsic barrier effect of the erythrocyte membrane, and the implications of these barriers on NO scavenging at differing physiological and pathophysiological hematocrits and oxygen tensions, we have studied NO scavenging rates of red cells using a number of independent methodologies. These studies confirm the important contributions of both the extracellular unstirred layer and, to a lesser extent, the intrinsic submembrane diffusional barrier. Interestingly, the latter effect increases with red cell oxygenation.
Stopped-flow mixing experiments were performed on an OLIS RSM 1000 spectrometer (Bogart, GA) coupled to a Molecular Kinetics threesyringe mixer (Indianapolis, IN). For anaerobic experiments, blood was deoxygenated by repeated cycling of gentle vacuum and atmospheric exchange with an inert, wet gas (nitrogen or argon, bubbled through water), and diluted into deoxygenated buffer, and residual oxygen was scavenged by sodium dithionite (0.2 mM final concentration after having been prepared in deoxygenated buffer under strict anaerobic conditions). For experiments on oxygenated cells, the samples were prepared in buffer that had been equilibrated in air. The diluted blood was loaded into a syringe in the Molecular Kinetics mixer so that the final concentration of Hb after stopped-flow mixing was 0.05 mM. The concentration of NO buffer, loaded into another syringe, after mixing was 2-11 times in molar excess to Hb. The third syringe was filled with buffer which, for experiments with deoxygenated cells, contained 0.5 mM dithionite that was used to scavenge oxygen in the mixing chamber and observation cuvette (1.5-mm path length), and the entire system was flushed constantly with argon. In initial anaerobic experiments, all three syringes contained 40% glucose by weight to reduce gravitational effects on the suspended red cells. To ensure that the glucose itself had no significant effect due its potential for osmotic activity, experiments were also carried out using an isotonic iodixanol solution, OptiPrep density gradient medium (25-30% by volume from Sigma) to reduce gravitational effects, and no difference was seen using this system. All experiments on oxygenated cells were performed using OptiPrep. The lack of any effect on tonicity of any of the buffers used was confirmed by incubating packed red cells diluted 5-fold in excess buffer and measuring the hematocrit percentage. The hematocrit values after incubation were 19% Ϯ 1%, 21% Ϯ 1%, and 17 Ϯ 2%, for the cells incubated in Dulbecco's phosphate-buffered saline alone, Dulbecco's plus OptiPrep, and Dulbecco's plus glucose, respectively. The viscosities of the solutions, measured using a glass capillary viscometer (Cannon Instruments, State College, PA), were 1.5 times that of water for those containing OptiPrep and 2.5 times that of water for those containing glucose. The observed rate of NO uptake by red cell-encapsulated Hb was determined by singular value decomposition (SVD) and global analysis (27) and divided by the final NO concentration to obtain an apparent bimolecular rate constant. The volumes and flow speeds were varied to examine potential artifacts on measured NO uptake rates, but the apparent bimolecular rates were found to be independent of these.
Competition experiments were performed similarly to those by the Liao group (11,28) with some modifications. NO reacts with deoxygenated Hb to form iron nitrosyl-Hb (HbNO, Fe(II)NO hemoglobin). NO reacts with oxygenated Hb to form methemoglobin (MetHb). NO was slowly released in excess cell-free and RBC-encapsulated Hb, and the reaction products (MetHb or HbNO) were examined to determine the relative rates of NO uptake by the RBCs compared with the reaction rate for cell-free Hb. Generally, red blood cells were washed in phosphatebuffered saline and resuspended at 15% or 50% Hct. Cell-free Hb was added to the RBCs suspension. The total amount of cell-free Hb reported is the sum of added Hb and that resulting from cell lysis during handling. The percentage of cell-free Hb to total Hb used varied between 0.4 and 3%. Previous work has shown that varying the amount of cell-free Hb while keeping the Hct constant does not significantly affect the results of competition experiments (11). A fresh ampoule of spermine NONOate was thawed and prepared in 0.01 M NaOH prior to addition to an Hb/RBC mixture. The ratio of cell-free to RBC-encapsulated Hb was determined by comparing the absorbance of the supernatant of the mixture after centrifugation at 7000 rpm for 2 min in a microcentrifuge to the absorbance of the diluted mixture. In some later experiments, the absorbance of the initial mixture was not taken, because it was calculated adequately based on the hematocrit and known additions of cell-free Hb. Absorption spectra were collected in the visible range (380 -700 nm) on a PerkinElmer Life Sciences lambda 9 spectrometer equipped with an integrating sphere to collect scattered light from the RBCs suspensions as described previously (26). Some spectra that did not contain RBCs were taken on a Cary 50 Bio spectrometer (Varian Inc.). The RBC/Hb mixtures were incubated with 10 -50 M NONOate while on a magnetic stirrer at slow speed to keep the sample homogeneous. Deoxygenated samples were also kept under a gentle nitrogen gas flow to maintain an anaerobic environment. Periodically, samples were collected to determine the amount of MetHb for experiments conducted under normoxic conditions or HbNO for experiments conducted under anaerobic conditions.
The amounts of MetHb and HbNO formed were determined by electron paramagnetic resonance (EPR). One portion (0.5 ml) of the mixture containing the total HbNO or MetHb formed was collected into a tube, and another portion was centrifuged (at 3000 ϫ g for 2 min). The supernatant containing only cell-free HbNO or MetHb was collected in another EPR tube, and a third EPR tube was filled with a sample containing the RBC pellet. HbNO was measured by EPR using a Bruker EMX 10/12 spectrometer operating at 9.4 GHz, 5-G modulation, 10.1milliwatt power, 655.36-ms time constant, and 167.77-s scan or 327.68-ms time constant and 83.89-s scan over 600 G at 110 K. MetHb was measured by EPR (at low field using 15-G modulation, 10.1-milliwatt power, 81.92-ms time constant, and 41.94-s scan over 700 G) at 4 K using liquid helium. The concentrations of Hb species measured by EPR were determined by performing the double integral calculation and comparing to standard samples.
The ratio of the bimolecular rate constant of NO reacting with cellfree Hb (k f ) to the apparent bimolecular rate constant of NO uptake by RBC-encapsulated Hb (k r ) was determined by the relative amount of HbNO or MetHb formed with, where the subscripts "r" and "f " refer to the RBCs encapsulated and cell-free Hb, respectively. The concentrations (indicated by brackets) represent the moles of the species in the total volume. Thus (for example), As long as the measured amounts of Hb that reacted with NO in the red cell and supernatant (cell-free) fractions added up to that measured in the un-fractionated portion, the three methods gave the same result. These three methods were therefore used as a self-consistency check for the amount of reacted Hb. Any data, where the percentage difference using one method of calculating k f /k r compared with the average of the three different methods was Ͼ30%, were thrown out. The equations we used to calculate k f /k r are similar to those derived previously (11). When necessary, a term accounting for the fact that [Hb] f is not constant in time due to conversion to HbNO or MetHb (11) was included.
In the experiments on oxygenated samples, problems could arise due to auto-oxidation of the Hb and activity of the RBCs MetHb reductase. To account for these processes, MetHb was measured on control samples that were not treated with the NO donor, and these were subtracted from the samples that were treated with the NO donor. Under the conditions of our experiments, MetHb reductase activity was not significant as evidenced by parallel experiments in which the amount of MetHb formed when NO donor was added to cell-free Hb alone was compared with the total MetHb formed when the NO was added to the mixture of cell-free Hb and RBCs. In these cases, the total amount of MetHb made in the samples without RBCs was always the same or less than that made in the ones with RBCs. The observed weak effect of the MetHb reductase system observed over the course of our experiments is consistent with the time course of MetHb decay following NO breathing (29).
For anaerobic experiments, samples were deoxygenated by gentle rocking and exposure to flow of wet argon or nitrogen gas overnight or by washing several times in deoxygenated buffer. For most experiments, sodium dithionite at a final concentration of 10 mM was used to scavenge residual oxygen. To test whether the dithionite itself had an effect, three separate experiments were conducted using a different oxygen scavenging system consisting of protocatechuate 3,4-dioxygenase at a final concentration of 0.05 units/ml and its substrate, protocatechuic acid at a final concentration of 1.0 mM. These concentrations did not significantly affect the pH. In addition, two experiments were conducted where no oxygen scavenging system was used, but extra care was taken in the handling of the sample to keep the Hb mostly deoxygenated. The only affect of the dithionite that was observed was that the rate of NO release was accelerated. No effect was observed on measured ratios of k f /k r .

RESULTS
Typical results from stopped-flow mixing of deoxygenated RBCs with excess NO are shown in Fig. 1. Fig. 1A shows the absorption spectra collected as a function of time over 1.8 s when 50 M hemoglobin (hematocrit of 0.25%) was mixed with 350 M NO under anaerobic conditions. Analysis of these spectra using singular value decomposition/ global analysis demonstrates that the spectral changes follow a single exponential and are indicative of the formation of HbNO (Fig. 1B). The initial spectrum is a mixture of deoxygenated Hb and HbNO. The formation of HbNO during the mixing time of the experiment is likely due to the presence of cell-free Hb as a result of limited hemolysis during mixing. As cell-free Hb binds NO governed by a rate constant of 3-6 ϫ 10 7 M Ϫ1 s Ϫ1 (30,31), the binding of NO to free Hb in this system would be expected to occur with a lifetime (1/rate) of Ͻ0.1 ms, and so it would occur within the dead time of mixing. The effective second-order rate constant for the binding of NO to RBC-encapsulated Hb was found to be (1.4 Ϯ 0.3) ϫ 10 4 M Ϫ1 s Ϫ1 (n ϭ 14 at 5 different NO:Hb ratios). In most experiments, glucose was used to prevent sedimentation of cells during the experiments. In one experiment, OptiPrep was used instead of glucose and the binding rate constant was measured to be (1.8 Ϯ 0.1) ϫ 10 4 M Ϫ1 s Ϫ1 . These values for NO binding to RBC-encapsulated Hb at low Hct (0.25%) and excess NO to Hb are about 1000 times slower than the measured rate of NO binding to cell-free Hb and confirm previous measurements (32).
Measurements on the reaction of excess NO to oxygenated RBCs are complicated by the fact that the NO will initially react with oxygenated Hb (which has absorption peaks at 576, 540, and 415 nm (33)) to form MetHb (which has absorption peaks at 500 and 405 nm (33)), and subsequently the MetHb can bind excess NO to form a ferric-NO species, MetHb-NO (which has absorption peaks at 564, 534, and 418 nm (34)). We present results from experiments on rapid mixing of NO and oxygenated RBCs in Fig. 2. Fig. 2 (A and B) shows species derived from SVD and global analysis of typical data collected with NO:Hb ratios of 2.4 ( Fig. 2A) and 6.4 (Fig. 2B). Fig. 2 (C and D) shows the corresponding time courses. Because MetHb has a relatively weak affinity for NO, when NO is not in large excess, very little MetHb-NO forms so that the reaction can be approximated as HbO 2 ϩ NO 3 MetHb governed by a single pseudo first order rate constant. This is the case for data such as that present in Fig. 2A, where we obtained an average observed rate of 2.6 s Ϫ1 . When NO is in greater excess to Hb, an additional reaction must be considered, MetHb ϩ NO 7 MetHb-NO. The situation is illustrated in We performed competition experiments under conditions where, unlike in stopped-flow experiments, Hb is in large excess to NO. For aerobic experiments, spermine NONOate was added to an oxygenated solution containing both RBCs and cell-free Hb, and the ratio of NO reacting with cell-free versus encapsulated Hb was measured by MetHb formation and used to calculate the relative rate constants (see "Materials and Methods"). For anaerobic experiments, the NO donor was added to deoxygenated RBCs, and cell-free Hb and HbNO was measured. Typical absorption and EPR spectra from these experiments are shown in Fig. 3. Fig. 3A shows an absorption spectrum from the supernatant of the oxygenated sample taken 67 min after the addition of spermine NONOate. This spectrum, together with known Hct, were used to calculate [Hb] f /[Hb] r , which was 0.01 for the data shown. EPR spectra were taken of the whole incubation mixture and of the supernatant and red cell pellet after centrifugation (Fig. 3C) to determine the level of cell-free and RBC-encapsulated MetHb ([MetHb] f , and [MetHb] r ). The EPR spectra (shown corrected to the total volume of the mixture) show that the fraction of MetHb in the RBCs plus that in the supernatant (containing only cell-free Hb) add up to the total MetHb measured in the sample that was not fractionated. Similar data are shown for an anaerobic sample in Fig. 3 (B and D). Fig. 4 shows the calculated concentrations of total, red cell, and cellfree MetHb or HbNO for samples collected at different times after addition of spermine NONOate for experiments at high (50%) and low (15%) Hct. These values of MetHb and HbNO were used to calculate the relative rates of NO uptake by cell-free and RBC-encapsulated Hb, k f /k r , using Equations 1 and 2. Whereas most of the reaction products are found in the supernatant fractions containing the cell-free Hb for aerobic experiments (Fig. 4, A and C), most are in the RBCs fractions for anaerobic experiments (Fig. 4, B and D), indicating that the hypoxic RBCs take up NO faster than the oxygenated ones. The ratio of cell-free Hb to RBC-encapsulated Hb, [Hb] f /[Hb] r , for the data shown for oxygenated samples at 15% Hct (Fig. 4A) was 0.004, and it was 0.01 for the data shown at 50% Hct (Fig. 4C). Based on Equations 1 and 2, if k f /k r is independent of Hct, one would have expected relatively more MetHb to have formed in the red cell fraction for the data shown at 15% Hct, but that was not the case. As it has been shown previously that k f /k r , does not depend on [Hb] f /[Hb] r (11), the large difference in k f /k r between the data shown at 50% and that shown at 15% (410 at high Hct versus 140 at low Hct) must be due to the Hct.
One may argue that the faster NO scavenging by deoxygenated cells is due to entry of the NONOate into the RBCs. To test for this, we split samples into two portions after 15 min of incubation with NONOate. One portion of the sample was spun down, and the supernatant was removed. The spun RBCs would only have NONOate that entered the RBCs. These cells were subsequently frozen at the 15-, 30-, 45-, and 60-min marks and assayed for HbNO formation. In one of these experiments, the amount of HbNO in the spun RBCs was found to be 16.1 M at all time points, whereas the portion of the sample that was not spun (containing extracellular NONOate) grew steadily from 15 to 43 M. In a repeat of the experiment, the largest deviation of the measured HbNO in the spun portion from the original amount was a 4% decrease (15.8 were rapidly mixed using a stopped-flow spectrophotometer, and absorption spectra were collected every millisecond over 1.8 s after mixing. Selected spectra, 37 ms apart, are illustrated. B, initial (labeled DeoxyHb) and final species (labeled HbNO) obtained from SVD and global analysis of the data from Fig. 1A. The data were fit to a single exponential process (A3 B). The initial spectrum actually contains some HbNO due to rapid NO binding to cell-free Hb during mixing. Inset, the time course of the measured absorbance at 430 nm plotted along with the fit obtained from SVD and global analysis. Initial time (zero) corresponds to the time when turbulence due to rapid mixing was complete.
M measured at the 45-min mark compared with 16.5 measured at the 15-min mark, immediately after sedimentation). Again, the HbNO in the unspun portion of the sample in the repeat experiment grew from 16.2 M at 15 min to 41.9 M at 60 min. These data prove that the NONOate does not enter the RBCs.
The calculated values of k f /k r from experiments performed under both oxygenated and deoxygenated conditions at 15 and 50% Hct are shown in Fig. 5. Our results from experiments performed under aerobic conditions at 15% Hct (410 Ϯ 93, n ϭ 8, 4 separate experiments) are consistent with those published previously (11), but we find a much lower k f /k r ratio under anaerobic conditions with cells at 15% Hct (k f /k r ϭ 51 Ϯ 19, n ϭ 19, 9 separate experiments). 4 Most of the anaerobic experiments were carried out using sodium dithionite to scavenge oxygen. In three experiments, when a different oxygen scavenging system was used (protocatechuate 3,4-dioxygenase/protocatechuatic acid), k f /k r was found to be 47 Ϯ 10 (n ϭ 8). When no oxygen-scavenging system was used (wherein the hemoglobin oxygen saturation was measured to be 10 Ϯ 5% and 20 Ϯ 10%), k f /k r was found to be 33 Ϯ 13 (two experiments, n ϭ 5). Thus, we find that deoxygenated red cells scavenge NO much faster than oxygenated ones.
At high (50%) Hct we found an average value of k f /k r for oxygenated cells of 140 Ϯ 44 (Fig. 5, n ϭ 7, 3 separate experiments). This is significantly smaller than the value we found for oxygenated cells at 15% Hct. For experiments performed under anaerobic conditions at 50% Hct, we found k f /k r ϭ 53 Ϯ 32 (Fig. 5, n ϭ 9, 4 separate experiments). This is not significantly different from the value found at 15% Hct for deoxygenated cells but is still smaller than the value found for oxygenated cells at 50% Hct.

DISCUSSION
In stopped-flow experiments on deoxygenated cells, conducted at low hematocrit and excess NO to Hb, we found that NO reacts with cell-free Hb ϳ1000 times faster than RBC-encapsulated Hb. This result is consistent with previous studies (11,13,32). To our knowledge, we are the first to report the results of measurements of NO uptake by oxygenated RBCs using stopped-flow absorption. We found that the second order rate constant for MetHb formation under conditions where subsequent formation of MetHb-NO could be ignored was (3.6 Ϯ 1.2) ϫ 10 4 M Ϫ1 s Ϫ1 . This is similar to the rate constant found from stopped-flow 4 As described under "Materials and Methods," any data where the percentage difference using one method of calculating k f /k r compared to the average of the three different methods was Ͼ30% was thrown out. This self-consistency test discards data where Hb from one fraction (red cell or cell-free) may contaminate another or data where the sum of the parts do not add up to the total Hb due to other artifacts. When this criterion was not applied, and all the data was included in the calculations, we found k f /k r for experiments at 50% Hct to be 190 Ϯ 140 for aerobic conditions and 56 Ϯ 60 for anaerobic conditions. When this criterion was not applied, and all the data was included in the calculations, we found k f /k r for experiments for low Hct to be 430 Ϯ 150 for aerobic conditions and 79 Ϯ 100 for anaerobic conditions.   in RBCs occurred at about the same rate as one would expect for its formation with cell-free Hb. This demonstrates (as discussed further below) that for slow bimolecular Hb reactions, there is no difference between the rate observed for RBCs and cell-free Hb (37). Together, our results using stopped-flow absorption confirm previous ones that NO uptake in these experiments is rate-limited by diffusion. Competition experiments could potentially uncover a role for intrinsic factors of the RBCs such as a physical membrane barrier in limiting the rate of NO uptake. Using competition experiments under aerobic conditions where Hb was in excess to NO we found that cell-free Hb scavenges NO about 400 times faster than RBC-encapsulated Hb at 15% Hct, consistent with previous results (11). Based on the previous result, it was suggested that most of the decreased NO scavenging by RBCs is due primarily to a physical membrane barrier (11). However, we found that, as the Hct is increased from 15% to 50%, the rate that NO is scavenged by oxygenated RBCs increases about 3-fold. This Hct dependence indicates that a large part of the 400-fold difference is actually due to external factors due to rate-limiting diffusion of NO (like an unstirred layer). In the competition experiments on deoxygenated cells, we observed fast NO scavenging by deoxygenated RBCs at 15% Hct, such that cell-free Hb only has a 50-fold advantage over RBC-encapsulated Hb. We suggest that the observed faster NO scavenging rate by deoxygenated cells is partially due to the fact that the bimolecular rate constant for NO binding to deoxygenated Hb is slower than the one for its reaction with oxygenated Hb. In addition, differences in the membrane barrier may be partially responsible for the observed differences between oxygenated and deoxygenated cells.
In stopped-flow experiments, where NO is in excess to Hb, both the unstirred layer and intrinsic membrane boundary can contribute to reduced NO uptake by RBCs. Liao and coworkers (11) developed a competition assay designed to reduce the contribution of the unstirred layer and thereby test its importance. It was argued that by using a slow NO donor, the concentration of NO would be homogeneous in the competition experiments and there would be no unstirred layer (11). This argument has been challenged by some investigators who hold that, even in the competition experiments, there is still an unstirred layer (13,25). Our demonstration of an Hct dependence on NO uptake by oxygenated RBCs supports the idea that external diffusion of NO, such as that set up in an unstirred layer, still plays a role in competition experiments. In the competition experiments, diffusion limits the rate of NO uptake by RBCs when the NO is released a large enough distance from the RBCs. However, the effects are minimized compared with stopped-flow experiments, partially because the rate of NO uptake in stopped-flow experiments is limited by the time it takes NO to diffuse a much further distance than in the competition experiments. In stopped-flow, the total concentration of NO is greater than that of Hb, but the local concentration of NO is less than that of hemoglobin in the RBCs (which is on the order of 20 mM). Thus, to saturate the RBCencapsulated Hb with NO, NO must diffuse very far in the stopped-flow experiments, which limits the overall kinetics. In the competition experiments, the NO donor concentration is always homogeneous, so rapid uptake by the RBCs is facilitated. The situation is similar to stoppedflow measurements of the initial rate of NO uptake, when the concentration of NO is (for a short time) homogeneous. The initial rate of NO uptake is much faster than rates calculated based on complete conversion of deoxygenated Hb to HbNO. Carlsen and Comroe (32) used stopped-flow absorption to measure the rate that NO is taken up by RBCs. They found the initial rate of NO uptake by RBC-encapsulated Hb to be between 1 ϫ 10 5 and 2 ϫ 10 5 M Ϫ1 s Ϫ1 (32). This is only about 100 times slower than the rate that cell-free Hb reacts with NO. Thus, competition experiments decrease the role of external diffusion in limiting NO uptake by RBCs, but our demonstrated Hct dependence shows that external diffusion still plays a significant role.
We observed a large difference between the relative NO uptake rates of oxygenated and deoxygenated cells: 8-fold at 15% Hct and almost 3-fold at 50% Hct. This difference is partially due to the difference in bimolecular rate constants for the reactions of oxygenated and deoxygenated cell-free Hb with NO (4 -8 ϫ 10 7 M Ϫ1 s Ϫ1 versus 3-6 ϫ 10 7 M Ϫ1 s Ϫ1 (31, 38 -41)). In competition experiments, scavenging of NO by RBC-encapsulated Hb is most rapid for NO produced close to the RBCs and slower for NO produced further away. The faster NO is scavenged outside of the RBCs, the less likely NO produced a given distance from the RBCs will be able to diffuse to it before being scavenged by cell-free Hb. Thus, one expects that, in the competition experiments, the faster the bimolecular rate constant for the NO/Hb reaction, the larger k f /k r will be. This has been shown computationally (25) and pointed out earlier for the case of hemoglobin ligands that combine with slower bimolecular rate constants, carbon monoxide, and isocyanide (37). The effect of the slower bimolecular rate constant for the NO/deoxygenated Hb reaction may also explain why we do not observe a significant Hct dependence in k f /k r for deoxygenated cells.
Although the differences observed in k f /k r for deoxygenated and oxygenated cells can be explained partially by differences in the bimolecular rate constants, the magnitude of the difference (3-to 8-fold) is also likely due to factors intrinsic to the RBCs such as a physical membrane barrier. The bimolecular rate constant for the reaction of NO with oxygenated Hb is only about 1.5 times that for deoxygenated Hb (26,31), and this is not likely to fully account for the up to 8-fold difference in k f /k r (25). The role for a physical membrane barrier is consistent with previous results showing chemical (14) or physical (15) modification of the red cell membrane cytoskeleton increased the rate of NO uptake. However, the effects of these modifications never produced a change in k f /k r greater than a factor of two. Thus, we conclude that the physical membrane barrier plays a role in limiting NO uptake, but not as large a role as previously proposed (11). Moreover, the membrane barrier is more effective in limiting NO uptake for oxygenated cells. One may speculate that this is due to binding of deoxygenated Hb to Band 3 tetramers in the cytoskeleton thereby displacing ankyrin, similar to a previously proposed mechanism involving intracellular HbNO modulation of RBCs NO uptake (28).
In 1999, Liao and coworkers (10) made the fundamentally important observation that the ability of RBCs to scavenge NO was decreased by flow. They also found that, in the absence of flow, to get an effect equivalent to that of 10 M free Hb, one needed about 1000 times more RBC-encapsulated Hb (about 50% Hct) (10). Thus, Liao and coworkers concluded that RBCs scavenge NO about 1000 times less efficiently than free Hb, even in the absence of a cell-free zone (10). This result is difficult to reconcile with our findings at 50% Hct. However, the scavenging of endothelial-derived nitric oxide by hemoglobin in isolated vessels is subject to a number of more complex considerations than are present in simple systems. RBC-endothelial interactions, even in the absence of flow, need to be considered, in addition to the extravasation of cell-free Hb into the subendothelial space. This latter process is thought to play a significant role in NO scavenging as indicated in studies in the development of blood substitutes (42)(43)(44)(45). It has been shown that the hypertensive effect of various hemoglobins is dependent on their size (43,44). Some of this effect may be due to the propensity of the hemoglobins to leave a cell-free layer, but as extravasated Hb can be found in the tissue (44), a role for extravasation is likely. At low concentrations of Hb a significant fraction of dimers will form, which can penetrate the endothelial glomerulus, and the remaining Hb tetramers are likely to extravasate through larger pores (44,46). It is possible that extravasation in vessel bioassays plays an especially significant role, because free Hb is known to increase Hb permeability itself due to its interaction with the endothelium (45). Thus, in the absence of flow, extravasation of free Hb may contribute significantly to enhanced NO scavenging by free Hb compared with RBCs in addition to any intrinsic differences.
Our results have important implications on hemodynamics in normal physiology and several diseased states. We have shown that the intrinsic rate of NO scavenging by oxygenated RBCs is highly dependent on Hct. It has long been known that the Hct of smaller vessels is less than that of larger vessels (47,48). This phenomenon, known as the Fåhraeus effect, is due to migration of red blood cells into the center of the vessels creating a cell-free zone near the endothelium layer, which results in a lower tube Hct (fractional volume of red cells in the vessel, including the cell-free zone) (47). The effect has been verified both experimentally and theoretically (47,48). The Hct dependence of NO scavenging we report here is likely to play a role in how NO is scavenged in different vessels in normal physiology. This surprising Hct dependence does not affect the notion that free Hb is a much more efficient NO scavenger than RBC-encapsulated Hb. Rather, our results taken together with previous ones, indicate that more of the enhanced NO scavenging stems from the cell-free zone and extravasation compared with an intrinsic reduced scavenging rate than previously thought. In any case, free Hb remains an efficient NO scavenger, and the goal of creating blood substitutes that have reduced NO scavenging (49) remains vital. In addition, the contribution of plasma cell-free hemoglobin released during intravascular hemolysis could be particularly important in hemolytic anemias, both as a consequence of the reduced rate of uptake of NO by red cells at low hematocrit and the importance of extravasation of free hemoglobin described above (23).